Validation of the Activpal Activity Monitor in Measurement of Total Steps Taken at Pre-Determined Low Intensity Walking Speeds

Guthrie, Kanoun

British Association of Sport and Exercise Sciences (BASES) Annual Student Conference, Heriot-Watt University, Edinburgh 30 – 31 March2006

It is well established that accurate measurement of Physical Activity (PA) is important but problematic. However, recent advances in technology have allowed for objective measurement using pedometery and accelerometery, which measure total steps taken as an indication of PA. Research investigating the validity of these devices has focused on walking speeds of ?54m/min and suggests that they have acceptable accuracy at speeds of ?67m/min (Bassett et al., 1996: Medicine and Science in Sports and Exercise, 28, 1071-1077; Crouter et al., 2003: Medicine and Science in Sports and Exercise, 35, 1455-1460). This indicates that these devices would be appropriate for populations that walk at these speeds. However, the accuracy of these devices is questionable for elderly individuals and patients who have been shown to walk at speeds <54m/min (Cyarto et al., 2004: Medicine and Science in Sports and Exercise, 36, 205-209). ActivPAL? professional physical activity monitor (PAL Technologies Ltd Glasgow) is a newly developed device, which has the potential to overcome these difficulties with its advanced mechanism of step detection and classification of posture. Therefore, this study aimed to establish criterion-related validity of the activPAL? activity monitor at pre-selected slow speeds. With institutional ethics approval forty-two participants (9 males, 33 females; mean age 23.5 4.0; mean BMI 22.0 ?22.0) provided written, informed consent and performed four 5-mintue periods of treadmill walking on the Woodway treadmill ELG 55 at each of the following speeds 27, 40, 54, 80m/min. Total steps taken during each 5-minute walking period was measured simultaneously by the activPAL? and direct observation using a hand-tally counter (the criterion measure). Following data collection, mean percentage accuracy of the activPAL? at each speed was calculated. Using the method recommended by Bland and Altman (1986: Lancet, 307-310) plots were created for each speed to assess the agreement between observed steps and activPAL? steps taken. This will demonstrate the bias (mean difference) and upper and lower limits of agreement (mean?1.96SD). The results in Table 1 show that the percentage accuracy and agreement of activPAL? with the criterion measure decrease as speed decreases. However, accuracy remains greater than 95%. The results therefore suggest that activPAL? has criterion-related validity at the slow speeds investigated, and imply that the device could be of use in populations who walk at a slow pace. However, further investigation of the activPAL? with populations who naturally ambulate at slow speeds, such as the elderly is required since they may possess certain gait abnormalities, which could affect device accuracy.

NA

Guthrie, Kanoun Validation of the Activpal Activity Monitor in Measurement of Total Steps Taken at Pre-Determined Low Intensity Walking Speeds British Association of Sport and Exercise Sciences (BASES) Annual Student Conference, Heriot-Watt University, Edinburgh 30 – 31 March2006 2006. NA.